Increased risk of neovascular age-related macular degeneration in patients with herpes zoster ophthalmicus: a retrospective cohort study.

Abstract

Herpes zoster is identified as a reactivation of latent varicella zoster virus (VZV) within a single dorsal root or cranial sensory ganglion (Oxman 2009). When the latent VZV reactivates in the ophthalmic branch of the trigeminal nerve, herpes zoster ophthalmicus (HZO) occurs (Vrcek et al. 2017). Of those patients with HZO, about 70% may develop ocular complications (Weinberg 2007). This cohort study aimed to explore the relationship between HZO and the subsequent risk of neovascular AMD using a population-based data set in Taiwan. Medical records for this cohort study were retrieved from the Taiwan Longitudinal Health Insurance Database 2005 (LHID2005) (TMU-JIRB N201705073). For the study cohort, we initially identified 1668 patients who had received a first-time principal diagnosis of HZO (ICD-9-CM codes 053.2) during an ambulatory care visit between 2001 and 2010. The date a patient received their first HZO diagnosis was identified as the index date. We then excluded 499 patients who were aged less than 40 years. In addition, we excluded 21 patients who had a medical history of dry AMD (ICD-9-CM code 362.51) or neovascular AMD (ICD-9-CM codes 362.52) prior to the index date. Finally, 1148 patients with HZO were recruited in the study cohort. The matched comparison cohort (n = 5740) (five comparison subjects per HZO patient) was selected from the remaining beneficiaries of the LHID2005. The comparison cohort was identified by matching HZO patients in terms of propensity score. The matching variables included patients’ sex, age, monthly income, urbanization level, geographical region, hypertension, coronary heart disease, hyperlipidaemia, tobacco use disorder and diabetes. The date of a comparison patient's first use of ambulatory care during that matched year was defined as the index date. Additionally, we ensured that none of the selected comparison subjects had been diagnosed with neovascular AMD before the index date. Each patient was individually tracked for a 3-year period starting from their index date to define those who were subsequently diagnosed with neovascular AMD. Table 1 shows that among all sampled patients, the incidence rate of neovascular AMD per 1000 person-years within the 3-year follow-up period was 2.31 (95% CI: 1.69–3.08). In addition, the findings indicated that incidence rates of neovascular AMD per 1000 person-years within the 3-year follow-up period were 6.82 (95% CI: 4.27–10.32) for HZO patients and 1.44 (95% CI: 0.92–2.13) for comparison subjects. The stratified Cox proportional analysis (stratified by propensity score) revealed that the hazard ratio of neovascular AMD during the 3-year follow-up period was 4.62 (95% CI: 2.59–8.24) for HZO patients compared to comparison subjects. This population-based cohort study found that HZO patients were 4.62 times more likely to suffer from neovascular AMD compared to those patients without HZO. Even though the actual pathogenesis of the association between HZO and neovascular AMD is uncertain, abnormal inflammation in HZO patients may be one of the possible mechanisms to explain this connection. The previous literature indicated that HZO patients usually also had ocular inflammation (Liesegang 2008; Sanjay et al. 2011). Some studies also reported that VZV infection via ophthalmic division of the trigeminal nerve may trigger a local inflammatory response that can further affect the retina, optic nerve, cornea, conjunctiva, iris (Bandeira et al. 2016; Vrcek et al. 2017), etc. Furthermore, the vasculopathy due to VZV infection might be another potential pathway to elucidate the relationship between HZO and subsequent neovascular AMD. This study revealed that HZO patients had a higher risk of subsequent neovascular AMD than matched comparison cohort. We suggest that further large epidemiological studies are warranted to ascertain the relationship between HZO and neovascular AMD in different populations.